AI News · Models ·
OpenAI says GPT-6.1 Sol Ultrafast is coming soon, without a release date
OpenAI product and engineering leader Tibo Sottiaux indicated on October 4 that GPT-6.1 Sol's faster mode is coming soon, RuntimeWire reported. His reply to developer Ben Davis followed the base model's September 29 launch. The promised Ultrafast mode remains a separate release with no firm launch date or announced pricing.
Key points
- GPT-6.1 Sol is available, but its separate Ultrafast mode has no announced release date or pricing.
- Tibo Sottiaux said “6.1 coming soon” on October 4 in an exchange about Ultrafast.
- OpenAI claims up to eight times faster token generation in Codex, not an equivalent improvement for every workflow.
- Standard Sol API rates are $2 per million input tokens and $10 per million output tokens.
- OpenAI’s Ultrafast guide does not yet give a Sol configuration or price.
What happened: OpenAI’s promised Ultrafast mode for GPT-6.1 Sol is still coming, but the company has not announced a release date or pricing, RuntimeWire reported. OpenAI product and engineering leader Tibo Sottiaux replied “6.1 coming soon” to developer Ben Davis on October 4. Although the reply did not name Ultrafast, RuntimeWire said the surrounding exchange concerned that faster mode. The message pointed to a pending speed option, not the announcement of another model.
Background: OpenAI launched GPT-6.1 Sol on September 29 for ChatGPT Work, Codex and API users. The company describes Sol as offering performance near its flagship GPT-6 Astra at a lower cost for coding, computer use and other professional work. Codex is OpenAI’s software-engineering agent. At Sol’s launch, OpenAI said the Ultrafast option would arrive “in the coming days.” Sottiaux’s reply five days later reaffirmed that plan without narrowing the release window. The standard model is already available for teams to test.
The details: OpenAI said Ultrafast would deliver up to eight times faster token generation than standard speed in Codex. That figure is the company’s stated maximum, not a guarantee that every prompt or complete workflow will run eight times faster. Sol’s standard API pricing is $2 per million input tokens and $10 per million output tokens, with cached input costing $0.10 per million tokens. OpenAI’s API documentation also lists a 1.05-million-token context window. Those prices apply to standard Sol, not the unreleased Ultrafast mode.
Who it affects: The distinction matters for developers and business teams considering Sol for applications where response time is important. Sol’s current pitch is lower token costs for work that would otherwise use a more capable model. Ultrafast is a separate option for users willing to pay more to reduce waiting. However, OpenAI has not disclosed the Sol Ultrafast rate. Teams therefore cannot yet compare its price with standard Sol or assess whether the promised speed improvement justifies the additional expense.
What to watch: The remaining questions are availability, pricing and configuration. RuntimeWire reported that OpenAI’s Ultrafast guide explains the API speed tier and how to select it for GPT-6 Astra, but does not give a Sol Ultrafast configuration or price. For now, teams can evaluate the available Sol model. Plans that depend on its faster mode still lack a confirmed launch date and the information needed to budget for that option.
Our take
Teams considering Sol for latency-sensitive applications should distinguish the available model from the promised faster mode. Deployment plans should rely on confirmed availability and pricing.